In theory, we coolly assess risks and alter our behaviour in order to minimise them, which is why many have given up beef, to avoid CJD, or come off the Pill, to ensure against thrombosis. In practice, the evidence suggests that our behaviour is motivated by panic and innumeracy.
Figures suggest that the risk of contracting BSE/CJD is really very small. Last year, 48 people died in the UK, and between 15 and 20 have so far contracted the new variety of CJD believed to be directly connected with BSE. And what of those timid souls still eating lamb instead? The higher fat content of the lamb gives them a far greater chance of early death from heart disease than contracting anything from BSE-infected beef.
It doesn't matter much, in the end. At a conservative estimate, you have about a thousand times greater chance of getting Alzheimer's from the aluminium in the pot your vegetables were cooked in than getting CJD from the meat.
Similarly, many people are afraid of flying, yet few shudder at the prospect of getting into a car or crossing a road. The comparative fatality rates make sober reading. Only six people die each year in British air crashes. Road deaths account for 1,749
It's true that we get into cars far more often than we board planes, just as we go up and down stairs more often than we take a bath. Yet, if you repeat one type of low-risk behaviour often enough, the combined effect may add up to an unacceptably high risk. You need to know how many roads you have to cross before a car hits you.
For example, average consumption of beef is 121 grammes per person per week. That works out at roughly one CJD death per 8 million kilograms of meat, or, to put it another way, one death per 45,000 cows. If you died in a thunderstorm, to take a further example, you would have already had to make 30 million walks.
One death also equals
10 million flights on a jet aeroplane
75 million trains boarded
200 million baths taken
4 million bicycle trips
500 million staircases climbed
3 million acts of unprotected sex
200 million roads crossed
15 million car journeys
2 million cigarettes
We are now in a position to make some valid comparisons. Cycling, for instance, is half as risky as smoking one cigarette, and getting in the bath is as dicey as crossing the road. If we persist in having unprotected sex, and also have a cigarette afterwards each time, the odds are that lung cancer will get us before Aids does - death by two million cigarettes strikes before the three million sex acts get you.
To most of us, a one-in-a-million chance is indistinguishable from a one-in-10-million. "I don't think people have any idea of what a million is," says David Hart, a member of the General Insurance Board of the Institute of Actuaries. He mentions the exuberance of those who win pounds 1,000 on the lottery; it seems to be as great as if they'd won a million. "People always think they're paying too much for motor insurance, for example," Mr Hart claims. "They're unaware of the very substantial loading on the premium to allow for the small numbers of serious injuries."
This area of the very small risk of very serious accident is where we make our biggest misjudgements. We naturally fail to allow for highly unlikely events until they happen, but, when they do, we overreact. We stop eating beef, but we still fail to fasten our seat belts in cars; we stop taking high-oestrogen contraceptive pills (which have a one-in- a-million chance of leading to a fatal thrombosis) and risk pregnancy instead (with its one in 20,000 mortality during childbirth). We over- protect our children on reading of one child being murdered (a 1-in-3 million chance each year), but we do not constantly look up to see if a falling object is about to strike us dead (1-in-600,000). Our assessment of risk is dictated less by the true odds than by our emotional reaction. Aids and CJD are just not our preferred ways of dying. We think of heart attacks, road accidents and lung cancer.
Even the insurance industry falls into this trap. Until the mid-1970s, subsidence was thrown in as a bonus to house policies. The droughts of 1975-76, however, forced them to reassess. Now there are areas of the country where you cannot find an insurance company willing to take your money.
The same mentality also explains our attitude towards positive risk, of winning the lottery or the football pools - or getting remarried ("A triumph of hope over experience," as Dr Johnson put it.).
In all cases, optimism overrules numeracy. And it is not only through underestimating large numbers. We may overestimate small numbers, too. Take the player who bets on red at roulette every time, doubling his bet when he loses. Like all good roulette strategies, it's bound to make a profit - as long as you don't run out of money first. And that would need 10 blacks in a row, which just never happens. Well, it does - about once a year if you play three games a night, as many penurious ex-gamblers and, in an analogous situation, Lloyds Names have found to their cost.
Looking on the bright side
Your lottery ticket will win at least something 2%
Your marriage will not end in divorce 67%
The weather forecast will be more or less correct 67%
You'll never be unemployed for more than a month 72%
Your car won't be broken into in the next year 80%
You won't die by accident 98%
Our perception of risk can also alter the risk itself. There is considerable evidence that wearing seat-belts makes drivers more reckless because they feel safer. Marking a road as an "accident black spot" may reduce accident figures so successfully that it ceases to be a black spot - it was only dangerous in the first place because people didn't realise how dangerous it was.
On the other hand, we may read that the roads are now safer for our children, a claim made recently on the basis of accident figures showing a dramatic decrease over the past two decades. But the roads are just as dangerous as they ever were. The difference is that parents are too frightened - of cars and other urban dangers - to let their children walk to school alone.
A few weeks ago, an official source in Mexico announced that the spewing of ash by Mount Popocatapetl posed no great threat to the 300,000 people living on its banks, or the two million within range of its lava flow, because the chance of a major eruption was less than one in 10. At the moment, it's still belching ash miles into the air, but only a few people have moved out of the area - not because one-in-10 is a bad risk, but because they can't stand the dust and constant Hoovering.
At the end of his excellent book, Risk (UCL Press, 1995), John Adams summarises his conclusions about risk management. One of the most striking points is this: "Safety interventions that do not alter people's propensity to take risks will be frustrated by responses that re-establish the level of risk with which people were originally content." In other words, however safe or dangerous the environment may be, we all settle for a level of risk-taking that keeps life as dangerous as we want it. As Marie Antoinette might have put it: "Let them eat sheep"
Sources: Mortality Statistics 1990; Road Accidents in GB 1995: The Casualty Report; General Household Survey 1995; Transport Statistics GB 1996; Annual Abstract of Statistics 1997; Social Trends 1997
Mortality figures per calendar year (or average figures for recent years): Deaths from: Lightning UK airline crash Train crash Drowning in bath CJD Cyclists Falling down stairs Aids Murder victim Pedestrians Car driver/passengers Lung cancer